Structural Basis for Spinophilin−Neurabin Receptor Interaction

regulation of synaptic transmission through their ability to target protein phosphatase 1 (PP1) to dendritic spines where PP1 dephosphorylates and ina...
0 downloads 0 Views 745KB Size
Biochemistry 2007, 46, 2333-2344

2333

Structural Basis for Spinophilin-Neurabin Receptor Interaction† Matthew S. Kelker,‡ Barbara Dancheck,‡ Tingting Ju,‡ Rene P. Kessler,‡ Jebecka Hudak,‡ Angus C. Nairn,§ and Wolfgang Peti*,‡ Department of Molecular Pharmacology, Physiology and Biotechnology, Brown UniVersity, 70 Ship Street, Box G-E3, ProVidence, Rhode Island 02912, and Department of Psychiatry, Yale UniVersity School of Medicine, New HaVen, Connecticut 06508 ReceiVed NoVember 13, 2006; ReVised Manuscript ReceiVed December 18, 2006

ABSTRACT: Neurabin and spinophilin are neuronal scaffolding proteins that play important roles in the regulation of synaptic transmission through their ability to target protein phosphatase 1 (PP1) to dendritic spines where PP1 dephosphorylates and inactivates glutamate receptors. However, thus far, it is still unknown how neurabin and spinophilin themselves are targeted to these membrane receptors. Spinophilin and neurabin contain a single PDZ domain, a common protein-protein interaction recognition motif, which are 86% identical in sequence. We report the structures of both the neurabin and spinophilin PDZ domains determined using biomolecular NMR spectroscopy. These proteins form the canonical PDZ domain fold. However, despite their high degree of sequence identity, there are distinct and significant structural differences between them, especially between the peptide binding pockets. Using two-dimensional 1H15N HSQC NMR analysis, we demonstrate that C-terminal peptide ligands derived from glutamatergic AMPA and NMDA receptors and cytosolic proteins directly and differentially bind spinophilin and neurabin PDZ domains. This peptide binding data also allowed us to classify the neurabin and spinophilin PDZ domains as the first identified neuronal hybrid class V PDZ domains, which are capable of binding both class I and II peptides. Finally, the ability to bind to glutamate receptor subunits suggests that the PDZ domains of neurabin and spinophilin are important for targeting PP1 to C-terminal phosphorylation sites in AMPA and NMDA receptor subunits.

Neurabin (1) and spinophilin (2-4) are neuronal scaffolding proteins that play important roles in synaptic transmission and synaptic plasticity (5, 6). Neurabin and spinophilin are highly enriched in dendritic spines, the site of excitatory neurotransmission. Neurabin is expressed almost exclusively in neuronal cells, while spinophilin is expressed ubiquitously, although it is highly enriched in neurons. Because spinophilin is a ubiquitous isoform of neurabin, it is sometimes termed neurabin II (3). Neurabin consists of 1095 residues (MW ) 122 730 Da), while spinophilin is smaller, consisting of only 817 residues (MW ) 89 640 Da). Figure 1a shows a domain representation for both proteins. As is typical for scaffolding proteins, both proteins contain multiple protein interaction domains. Both neurabin and spinophilin contain an F-actin binding, a PP1binding, a PDZ, and a C-terminal coiled-coil domain. In addition, neurabin, but not spinophilin [in vertebrates (7)], contains a sterile R motif (SAM) domain in its C-terminus, † This work was funded in part by a Richard B. Salomon Faculty Research Award, a Medical Research Grant of the Rhode Island Foundation, and Brown University Start-up Funds to W.P. M.S.K. is supported by NIH-NSRA Fellowship 1F32NS054493-01A1. This material is based upon work supported under a National Science Foundation Graduate Research Fellowship to B.D. R.P.K. is supported by a Karen T. Romer Undergraduate Teaching and Research Award. A.C.N. is supported by grant NIH-MH074866. The ITC instrument was purchased using NSF/EPSCoR funds (Award 0554548). * To whom correspondence should be addressed. Phone: (401) 8636084. Fax: (401) 863-6087. E-mail: [email protected]. ‡ Brown University. § Yale University School of Medicine.

while spinophilin, but not neurabin, is proposed to have a dopamine receptor-R-adrenergic interacting domain in its N-terminus, possibly between spinophilin residues 200 and 400 (8, 9). The highest level of primary sequence identity between neurabin and spinophilin is found in the PDZ domains (86%), the protein phosphatase 1 (PP1) binding domains (81%), and the coiled-coil domains (63%). Both neurabin and spinophilin have a central role in signaling in dendritic spines, as they function to target PP1 toward components of both glutamatergic (fast synaptic transmission) and dopaminergic (slow synaptic transmission) signaling pathways (10). Changes in the phosphorylation state of the postsynaptic glutamate receptor R-amino-3-hydroxy5-methyl-4-isoxazolepropionic acid (AMPA1)-type channels are important for synaptic plasticity. Regulation of N-methylD-aspartate (NMDA)-type glutamate receptors also contributes to control of synaptic transmission and plasticity. Protein kinase A (PKA) and calcium/calmodulin-dependent kinase II (CaMKII) play critical roles in phosphorylation and regulation of AMPA receptor trafficking and activity, with PP1 being able to reverse the action of these kinases through 1 Abbreviations: NMR, nuclear magnetic resonance; HSQC, heteronuclear single-quantum correlation; rmsd, root-mean-square deviation; AMPA, R-amino-3-hydroxy-5-methylisoxazole-4-propionic acid; NMDA, N-methyl-D-aspartic acid; PKA, protein kinase A; CaMKII, calcium/calmodulin-dependent kinase II; DARPP-32, dopamine- and cyclic AMP-regulated phosphoprotein with a relative molecular mass of 32 000 Da; NOE, nuclear Overhauser effect; ITC, isothermal titration calorimetry; PP1, protein phosphatase 1.

10.1021/bi602341c CCC: $37.00 © 2007 American Chemical Society Published on Web 02/06/2007

2334 Biochemistry, Vol. 46, No. 9, 2007

Kelker et al.

FIGURE 1: Primary amino acid sequences of spinophilin and neurabin. (a) Domain structure of spinophilin (top) and neurabin (bottom). Domains of spinophilin and neurabin studied in this work are highlighted with a black box. (b) Sequence alignment of the spinophilin and neurabin PDZ domains. Experimentally determined secondary structure assignments are annotated as cartoons. The consensus peptide binding sequence, GLGI, is highlighted with a black frame.

its ability to dephosphorylate AMPA receptors. PP1 is also able to dephosphorylate and regulate NMDA receptors (11). Through their ability to target PP1 to dendritic spines, neurabin and spinophilin bring the phosphatase into the proximity of AMPA and NMDA receptors. Moreover, localization at dendritic spines allows spatial control of PP1 activity by additional signaling pathways involving D1 dopamine receptor-dependent activation of PKA, activation of the PP1 inhibitor, DARPP-32 (dopamine- and cyclic AMP-regulated phosphoprotein with a relative molecular mass 32 000 Da), and PKA-dependent regulation of the interaction of spinophilin with the F-actin cytoskeleton within dendritic spines (12). PDZ domains, which were originally identified as conserved elements in the postsynaptic density protein PSD95, the disc-large tumor suppressor Dlg, and the Zonula occludens protein ZO-1, are modular protein interaction domains that recognize and bind C-terminal tetrapeptides from interaction proteins, such as transmembrane receptors (13, 14). They have a critical role in anchoring supramolecular signaling complexes to receptor proteins (15). Since PDZ domains mediate binding between large protein-protein assemblies that are involved in signaling and subcellular transport, they have received significant attention as potential drug targets for highly specific signaling pathway regulation (16). All PDZ domains have a conserved Gly-Leu-Gly-Φ binding motif (GLG-Φ, where Φ is a hydrophobic residue and most often is Phe) within the βA and βB connecting loop. This loop is coordinated via a hydrogen bond with the C-terminus of the binding peptide. The binding specificity of PDZ domains is significantly determined by the interaction of the first residue of helix RB and the side chain of residue

-2 of the C-terminal ligand. This interaction has resulted in the identification of five distinct classes of PDZ domains (13, 16, 17). (1) Class I PDZ domains bind peptides with the consensus sequence -X-S/T-X-Φ-COOH. The OH group of Ser/Thr forms a hydrogen bond with the class I conserved N3 atom of histidine in helix RB. (2) Class II PDZ domains bind the sequence -X-Φ-X-Φ-COOH (hydrophobic-hydrophobic interaction). (3) Class III PDZ domains bind the sequence -X-D/E-X-Φ-COOH, where the acidic residue coordinates with a class III conserved Tyr in helix RB. (4) Class IV PDZ domains bind the sequence -X-Ψ-D/E-COOH, where Ψ is an aromatic residue (18, 19). (5) Class V hybrid PDZ domains are capable of binding both class I type and class II type C-terminal peptides (20). Despite the significant improvement in the detailed physiological understanding of the spinophilin-PP1-DARPP32-AMPA signaling network (6, 21), extremely limited structural information about the protein interactions that regulate this system is available. Thus, we used biomolecular NMR spectroscopy to elucidate the three-dimensional (3D) structures of the spinophilin and neurabin PDZ domains. Furthermore, we used NMR titration experiments with C-terminal peptides derived from AMPA and NMDA receptors and literature-reported cytosolic interaction proteins to determine if these C-terminal peptides bind the spinophilin and/or neurabin PDZ domains. The results show that spinophilin and neurabin PDZ domains are the first known neuronal class V PDZ domains and that they interact directly with the C-termini of both AMPA and NMDA receptors, interactions that may contribute directly to the ability of PP1 to dephosphorylate closely adjacent C-terminal phosphorylation sites.

NMR Analysis of Neuronal PDZ Domains MATERIALS AND METHODS Protein Expression and Purification. Constructs representing spinophilin493-583, spinophilin493-602, or neurabin502-594 were subcloned into a vector derived from pET-28a (Novagen), which encodes a Thio6His6 expression/purification tag (MGSDKIHHHHHH) and a TEV (tobacco etch virus) cleavage site (ENLYFQGH) (22). The plasmids were transformed into Escherichia coli strain BL21-CodonPlus (DE3)-RIL (Stratagene). The expression of uniformly 13C- and 15N-labeled and 15N-labeled protein was carried out by growing freshly transformed cells in M9 minimal medium containing 4 g/L [13C]-D-glucose and/or 1 g/L 15NH4Cl as the sole carbon and nitrogen sources, respectively. Cell cultures were grown at 37 °C with vigorous shaking to an OD600 of 0.6-0.8. The expression of neurabin502-594, spinophilin493-583, and spinophilin493-602 was induced with 1 mM IPTG. The temperature was lowered to 18 °C, and the cell cultures were grown for 18 h. The cells were harvested by centrifugation, resuspended in extraction buffer [10 mM Tris-HCl (pH 8.0), 250 mM NaCl, 5 mM imidazole (pH 8.0), 0.1% Triton X-100, and Complete EDTA free tablets (Roche)], and lysed by cell cracking (Avestin C-3 Emulsiflex). The cell debris was removed by centrifugation (20000g for 20 min at 4 °C). For the first purification step, the soluble proteins were either loaded onto a HisTrap HP column (GE Healthcare), equilibrated with 50 mM Tris-HCl (pH 8.0), 5 mM imidazole, and 500 mM NaCl, or onto a Ni-NTA column (Invitrogen) pre-equilibrated with 10 mM Tris-HCl (pH 8.0), 250 mM NaCl, and 5 mM imidazole (pH 8.0). The protein was eluted with a 5 to 500 mM imidazole gradient. Fractions containing the proteins were pooled and exchanged into a buffer containing 50 mM Tris (pH 7.5) and 50 mM NaCl. After addition of TEV NIa (S219V) protease fused to an in-frame His6 tag, the solutions were incubated at room temperature for 1-3 days. The progression of TEV cleavage was monitored using SDS-PAGE analysis. After cleavage was at least 95% complete, the samples were exchanged into a buffer containing 10 mM Tris-HCl (pH 8.0), 250 mM NaCl, and 5 mM imidazole (pH 8.0) and loaded onto a preequilibrated Ni-NTA column (Invitrogen). The flow-through contained only cleaved neurabin502-594, spinophilin493-583, or spinophilin493-602 with an extra N-terminal GHM tripeptide (residues 1-3) as determined by SDS-PAGE. The solution was then concentrated to a final volume of 550 µL. The final concentrations for the different NMR samples were 2 mM for spinophilin493-583, 2 mM for spinophilin493-602, and 1.5-3 mM for neurabin502-594 [20 mM sodium phosphate (pH 6.5), 50 mM NaCl, and 10% D2O] (23). Most NMR measurements were performed at 298 K on a Bruker AvanceII 500 MHz spectrometer using a TCI HCN z-gradient cryoprobe. Some experiments for neurabin502-594 were conducted on a Bruker DRX 600 MHz spectrometer using a room-temperature TXI HCN z-gradient probe. Proton chemical shifts were referenced to internal 3-(trimethylsilyl)1-propanesulfonic acid, sodium salt (DSS). Using the absolute frequency ratios, the 13C and 15N chemical shifts were referenced indirectly to DSS. Chemical Shift Assignment and Structure Calculation. The following spectra were used to achieve the sequence-specific backbone and side chain assignments of all aliphatic residues: two-dimensional (2D) 1H-15N HSQC, 2D 1H-

Biochemistry, Vol. 46, No. 9, 2007 2335 13

C HSQC, 3D HNCACB, 3D CBCA(CO)NH, 3D CC(CO)NH, 3D HNCO, 3D HNCA, 3D HBHA(CO)NH, 3D 15Nresolved 1H-1H TOCSY, and 3D HC(C)H-TOCSY (24). The 2D 1H-1H NOESY, 2D 1H-1H TOCSY, and 2D 1H1 H COSY spectra of the spinophilin493-602 and neurabin502-594 samples in D2O solution after complete H-D exchange of the labile protons were used for the assignment of the aromatic side chains. The NMR spectra were processed with Topspin1.3 (Bruker, Billerica, MA) and analyzed with the CARA software package (www.nmr.ch). For the 3D structure determination of neurabin502-594 and spinophilin493-602, we used established NMR methods for the sequence-specific backbone and side chain assignments (24). Semiautomated programs were used to evaluate the data (CARA). We used the following spectra for the structure calculation: 3D 15N-resolved 1H-1H NOESY, 3D 13Cresolved 1H-1H NOESY (mixing time of 85 ms), and 2D 1 H-1H NOESY (mixing time of 85 ms, D2O solution). NOESY peak picking, NOESY peak assignment, and 3D structure calculation were performed automatically, using the ATNOS/CANDID/CYANA software package (25-27). The inputs for the structure calculations of the neurabin and spinophilin PDZ domains were amino acid sequences, the complete chemical shift lists, and the 3D and 2D NOESY spectra. Constraints for backbone dihedral angles derived from 13C chemical shifts were used only in the initial structure calculation. In a second step, the automatically picked NOESY peak lists were manually improved and used as input for the CYANA program package, which was used for the final structure calculation. This additional manual step improved the quality of the structures substantially. Our experience shows that the ATNOS/CANDID approach works optimally with high-field NOESY spectra recorded at 800 and 900 MHz because of an improved resolution and signal to noise ratio (28-30). However, when this additional manual step is performed, NOESY spectra recorded at lower field, like ours at 500 MHz, can also be used successfully. A total of 2217 NOESY-derived distance constraints (∼24 NOE constraints per residue) were used for the structure calculation of neurabin502-594 and 1507 (∼14 NOE constraints per residue) for spinophilin493-602. The decrease in the number of NOE constraints per residue for spinophilin compared to neurabin is easily ascribed to the 19-residue highly flexible C-terminal linker in spinophilin, for which very few NOE constraints were found and which increased the number of unresolved peaks in the 2D and 3D NOESY spectra (Table 1). The recently reported RECOORD scripts, in conjunction with CNS, were used for energy refinement in a water shell (31, 32). The neurabin502-594 model has excellent stereochemistry, with 98.6% of the residues in the most favored and additionally allowed regions of the Ramachandran diagram, 1% in the generously allowed region, and 0.4% in the disallowed region. Similarly, for spinophilin493-602, 97.5% of the residues are in the most favored and additionally allowed regions, 1.6% in the generously allowed region, and 0.9% in the disallowed region. Again, the lower stereochemical quality of the spinophilin PDZ domain structure, determined by a 1.1% increase in the percentage of residues in the generously allowed and disallowed regions of the Ramachandran plot, is due to increased spectral overlap and the lack of detectable long-range NOEs in the largely unstructured C-terminal

2336 Biochemistry, Vol. 46, No. 9, 2007

Kelker et al.

Table 1: Structural and CNS Refinement Statistics no. of restraints unambiguous distance restraints (all) intraresidual sequential medium-range long-range deviations from idealized covalent geometry bonds (Å) angles (deg) impropers (deg) structural quality Ramachandran plot [NMR-PROCHECK (34)] most favored region (%) additionally allowed region (%) generously allowed region (%) disallowed region (%) pairwise rmsd (Å) backbone (N, CR, C, and O) (4-25, 35-93) all heavy atoms (4-25, 35-93)

spinophilin493-602

neurabin502-594

1507

2217

437 471 208 391

474 618 378 747

0.012 ( 0.0003 1.32 ( 0.033 1.56 ( 0.11

0.013 ( 0.0005 1.33 ( 0.034 1.53 ( 0.09

73.0 24.5

77.4 21.2

1.6

1.0

0.9

0.4

0.88 ( 0.18

0.58 ( 0.11

1.41 ( 0.2

1.25 ( 0.17

Table 2: C-Terminal Peptides Derived from AMPA and NMDA Receptors and Literature-Reported Interaction Proteins receptor AMPA

protein

GluR1 GluR2 GluR3 GluR4 NMDA NR1C2 NR1C2′ NR2A/B NR2C/D NR3A NR3B

kexa

interaction surfaceb

peptide

PDZ class

GATGL ESVKI GTESVKI ASDLP SRHRES SVSTVV SIESDV SLESEV NRTCES AAPAES

I II II

.104 .104

small small

I I I

.104 .104 .104

large large large

.104 102-104 102-104

large small large

AM-NR hybrid ESVKV P70S6 Kinase EHLRMNL Kalirin-7 DPFSTYV

II II I

a kex is the estimated peptide-protein exchange constant (51, 52). Interaction surface correlates with the number of residues involved in complex formation (Figure 6).

b

linker tail of the spinophilin493-602 PDZ domain. The chemical shift assignment of the spinophilin PDZ domain has been recently published (23). The quality of the structures was assessed with WHATCHECK (33), AQUA (34), NMRPROCHECK (34), and MOLMOL (35). All structure comparisons throughout this work were performed using the closest conformer to the mean structure for all PDB entries. Peptide Library Screening. Peptide ligands derived from the C-terminus of NMDA and AMPA channels and cytosolic proteins [P70 S6 kinase (36) and kalirin-7 (37)] are listed in Table 2. Peptides were obtained at >95% purity either from JPT Peptide Technologies Inc. or from the Keck Foundation Biotechnology Resource Laboratory at Yale University. All peptides were solubilized in the same buffer that was used for the NMR measurements of the PDZ domains. The NR1C2′ peptide (SVSTVV) was the only peptide not soluble in this buffer and was solubilized by DMSO. DMSO without the NR1C2′ peptide was added to the neurabin and spinophilin PDZ domains and a 2D 1H-15N HSQC spectrum

recorded to ensure no interaction of DMSO with the PDZ domains (Figure S4 of the Supporting Information). Interaction study measurements were carried out on a Bruker AvanceII 500 MHz spectrometer equipped with a TCI HCN z-gradient cryoprobe. A 2D 1H-15N HSQC spectrum was used to monitor perturbations in 1H and 15N chemical shifts, which occur due to peptide binding. Comparison of the bound and unbound 2D 1H-15N HSQC spectra was used to detect binding using TopSpin1.3 (Bruker). Data were analyzed using the CARA software package (www.nmr.ch). We calculated dissociation constants (Kd) from the NMR chemical shift perturbation measurements. Chemical shift resonances of spinophilin/neurabin PDZ domains that exhibited the most significant change upon peptide interaction (GLGΦ binding motif, βB, RB) and, in addition, were well-resolved were used for these calculations. The averaged proton and nitrogen chemical shift changes [x(δH-δ0)2+(δN-δ0)2] were fitted to the following standard equation using the nonlinear regression analysis package of Mathematica 5.2 (Wolfram Research Inc.):

[m0 + p0 + Kd - x(m0 + p0 + Kd)2 - 4m0p0] δ ) δm 2m0 where m0 and p0 are the initial concentrations of the PDZ domain and the peptide, respectively, δm is the maximum chemical shift change for the protein-peptide complex (1:5 or 1:10 measurements), and Kd is the dissociation constant. In addition, we used isothermal titration calorimetry (ITC) measurements using a Microcal (Northampton) VP-ITC microcalorimeter to confirm the Kd values obtained using NMR data analysis (Table S1 of the Supporting Information). Since ITC measurements for protein-peptide interaction studies with Kd values in the low micromolar range, the typical range expected for PDZ-peptide interactions, require a large amount of protein and peptide, only a selective experiment was conducted to confirm the NMR titration results. Chemical Shift Assignments and Coordinates. Chemical shift assignments of neurabin502-594 were deposited in the BMRB as entry 6933, and coordinates were submitted to the Protein Data Bank as entry 2FN5. Chemical shift assignments of spinophilin493-602 were deposited in the BMRB as entry 6927, and coordinates were submitted to the Protein Data Bank as entry 2G5M. RESULTS Structural Analysis 3D Structures of the Neurabin and Spinophilin PDZ Domains. Two constructs for the spinophilin PDZ domain and one construct for the neurabin PDZ domain were expressed and purified. The first of these constructs is spinophilin493-583, which comprises the spinophilin PDZ domain, and the second is spinophilin493-602, which comprises the spinophilin PDZ domain and a 19-amino acid C-terminal extension that functions as a linker to the C-terminal coiledcoil domain (Figure 1A,B; Figure S1 of the Supporting Information). Secondary structure prediction programs (38) predict that this linker may form an R-helix. By comparison of the 2D 1H-15N HSQC spectra of both constructs, 16

NMR Analysis of Neuronal PDZ Domains

Biochemistry, Vol. 46, No. 9, 2007 2337

FIGURE 2: NMR structures of (a) neurabin502-594 and (b) spinophilin493-602 CR polypeptide backbone chains of a bundle of 20 energyminimized conformers superimposed for minimal rmsd values of the backbone atoms of residues 504-529 and 539-590 in neurabin and corresponding residues in spinophilin. The C-terminus of spinophilin493-602 is disordered (turquoise correlating to turquoise residues in Figure 1B). R-Helices are colored red, β-strands yellow, and all residues not in regular secondary structural elements green. Secondary structural elements are labeled following common PDZ domain nomenclature. The PDZ domain binding pocket is flanked by β-strand βB and helix RB. The numbers 502 and 594, and 483 and 602, identify the N- and C-termini for neurabin and spinophilin, respectively. This figure and all subsequent figures were generated using PyMOL (53).

additional cross-peaks were identified in the 2D 1H-15N HSQC spectrum of spinophilin493-602. These 16 additional cross-peaks correlate well with the 19-amino acid extension of this construct. Importantly, the chemical shift pattern of the spinophilin PDZ domain, identified by overlap of the 2D 1H-15N HSQC spectrum of spinophilin493-602 with that of spinophilin493-583, exhibited identical 1H-15N chemical shifts (except the last three C-terminal residues now extended by the 19-residue linker), demonstrating that the 3D structure of the spinophilin PDZ domain is not influenced by the C-terminal linker. Therefore, the structure of spinophilin493-602 was elucidated to determine whether there is residual structure in the linker region. In addition, the structure of the neurabin PDZ domain, neurabin502-594, without this C-terminal linker was elucidated, to characterize potential structural differences between the spinophilin and neurabin PDZ domains. The Extended C-Terminal Loop of Spinophilin Is Unstructured. As expected, the neurabin502-594 and spinophilin493-602 structures form typical PDZ folds (Figure 2). The 3D structures of the neurabin and spinophilin PDZ domains, like most PDZ domains, are comprised of six β-strands, βAβF, and two R-helices, RA and RB. These secondary

structure elements fold into a six-stranded β-sheet flanked by the two R-helices. The C-terminal recognition peptide binding groove is located between the βB strand and the RB helix. Notably, no secondary structure, based on chemical shift indexing and HN-HN NOE analysis, was determined in the C-terminal extension of spinophilin493-602. Instead, the 19residue C-terminal extension appears to be highly flexible, based on its increased intensity of NH cross-peaks (increased transverse relaxation time) in the 2D 1H-15N HSQC spectrum and dramatically reduced numbers of cross-peaks in the 3D 15N-resolved 1H-1H NOESY spectrum when compared to NH cross-peaks from the rest (core PDZ domain) of the protein. Outside of this linker, the most notable difference between the two PDZ domains is that the loop between βB and βC of spinophilin is angled toward and cradles the RB helix, whereas in the neurabin structure, it does not (Figures 2 and 3). There are five glycine and two alanine residues in the βB and βC loop region which likely contribute to its high conformational flexibility. Thus, it seems that the highly mobile 19-residue C-terminal tail of spinophilin, which occupies a large amount of steric space, constrains this loop to the observed more bent conformation.

2338 Biochemistry, Vol. 46, No. 9, 2007

FIGURE 3: Comparison of the 3D structures of the neurabin (green) and spinophilin (red) PDZ domains determined in this study. Core residues with a rmsd of